> Chris Gilbert wrote:

> I was wondering if some kind of heavy disk read/write etc might exercise irq 
> handling, but then you might find that it gets hidden by the vm system...
> 
How about serial or network io. According to systat vmstat serial IO 
gives an interrupt rate of 2000 interrupts per second + 2000 softserial 
interrupts a second. This is running pppd attached to a 56k modem. I'll 
give your patch a try at let you know how it goes.

> true, however I'd hope that the removal of the checking of every bit in the 
> irq request value would hide any issue like that. (and the ffs table is only 
> 64 bytes big, which is 2 cache lines on the SA)
> 
I have to admit I tried a similar patch to yours but I was never able 
to prove any kind of performance enhancement. 

> 
> It should be useable elsewhere, but I've not investigated the irq handling on 
> other systems yet.  I've only got access to a riscpc, but I do believe that 
> the core code will be similiar.
> 
The irq handling code for CATS and SHARK seems to be almost identical 
to that on the RISCPC.  It's probably not useful on an SA1110 based machine 
since the SA1110 has a hardware interrupt masking system which should 
be flexible enough to be exploited by NetBSD. IIRC the SA1100 has the 
same irq handling hardware as the SA1110 but its a long time since I 
looked at the datasheet.

Mike